The Synergy of LLMs & RL Unlocks Offline Learning of Generalizable Language-Conditioned Policies with Low-fidelity Data

Thomas Pouplin, Kasia Kobalczyk, Hao Sun, Mihaela Van Der Schaar
Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:49628-49658, 2025.

Abstract

Developing autonomous agents capable of performing complex, multi-step decision-making tasks specified in natural language remains a significant challenge, particularly in realistic settings where labeled data is scarce and real-time experimentation is impractical. Existing reinforcement learning (RL) approaches often struggle to generalize to unseen goals and states, limiting their applicability. In this paper, we introduce $\textit{TEDUO}$, a novel training pipeline for offline language-conditioned policy learning in symbolic environments. Unlike conventional methods, $\textit{TEDUO}$ operates on readily available, unlabeled datasets and addresses the challenge of generalization to previously unseen goals and states. Our approach harnesses large language models (LLMs) in a dual capacity: first, as automatization tools augmenting offline datasets with richer annotations, and second, as generalizable instruction-following agents. Empirical results demonstrate that $\textit{TEDUO}$ achieves data-efficient learning of robust language-conditioned policies, accomplishing tasks beyond the reach of conventional RL frameworks or out-of-the-box LLMs alone.

Cite this Paper

BibTeX
@InProceedings{pmlr-v267-pouplin25a, title = {The Synergy of {LLM}s & {RL} Unlocks Offline Learning of Generalizable Language-Conditioned Policies with Low-fidelity Data}, author = {Pouplin, Thomas and Kobalczyk, Kasia and Sun, Hao and Van Der Schaar, Mihaela}, booktitle = {Proceedings of the 42nd International Conference on Machine Learning}, pages = {49628--49658}, year = {2025}, editor = {Singh, Aarti and Fazel, Maryam and Hsu, Daniel and Lacoste-Julien, Simon and Berkenkamp, Felix and Maharaj, Tegan and Wagstaff, Kiri and Zhu, Jerry}, volume = {267}, series = {Proceedings of Machine Learning Research}, month = {13--19 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v267/main/assets/pouplin25a/pouplin25a.pdf}, url = {https://proceedings.mlr.press/v267/pouplin25a.html}, abstract = {Developing autonomous agents capable of performing complex, multi-step decision-making tasks specified in natural language remains a significant challenge, particularly in realistic settings where labeled data is scarce and real-time experimentation is impractical. Existing reinforcement learning (RL) approaches often struggle to generalize to unseen goals and states, limiting their applicability. In this paper, we introduce $\textit{TEDUO}$, a novel training pipeline for offline language-conditioned policy learning in symbolic environments. Unlike conventional methods, $\textit{TEDUO}$ operates on readily available, unlabeled datasets and addresses the challenge of generalization to previously unseen goals and states. Our approach harnesses large language models (LLMs) in a dual capacity: first, as automatization tools augmenting offline datasets with richer annotations, and second, as generalizable instruction-following agents. Empirical results demonstrate that $\textit{TEDUO}$ achieves data-efficient learning of robust language-conditioned policies, accomplishing tasks beyond the reach of conventional RL frameworks or out-of-the-box LLMs alone.} }
Endnote
%0 Conference Paper %T The Synergy of LLMs & RL Unlocks Offline Learning of Generalizable Language-Conditioned Policies with Low-fidelity Data %A Thomas Pouplin %A Kasia Kobalczyk %A Hao Sun %A Mihaela Van Der Schaar %B Proceedings of the 42nd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2025 %E Aarti Singh %E Maryam Fazel %E Daniel Hsu %E Simon Lacoste-Julien %E Felix Berkenkamp %E Tegan Maharaj %E Kiri Wagstaff %E Jerry Zhu %F pmlr-v267-pouplin25a %I PMLR %P 49628--49658 %U https://proceedings.mlr.press/v267/pouplin25a.html %V 267 %X Developing autonomous agents capable of performing complex, multi-step decision-making tasks specified in natural language remains a significant challenge, particularly in realistic settings where labeled data is scarce and real-time experimentation is impractical. Existing reinforcement learning (RL) approaches often struggle to generalize to unseen goals and states, limiting their applicability. In this paper, we introduce $\textit{TEDUO}$, a novel training pipeline for offline language-conditioned policy learning in symbolic environments. Unlike conventional methods, $\textit{TEDUO}$ operates on readily available, unlabeled datasets and addresses the challenge of generalization to previously unseen goals and states. Our approach harnesses large language models (LLMs) in a dual capacity: first, as automatization tools augmenting offline datasets with richer annotations, and second, as generalizable instruction-following agents. Empirical results demonstrate that $\textit{TEDUO}$ achieves data-efficient learning of robust language-conditioned policies, accomplishing tasks beyond the reach of conventional RL frameworks or out-of-the-box LLMs alone.
APA
Pouplin, T., Kobalczyk, K., Sun, H. & Van Der Schaar, M.. (2025). The Synergy of LLMs & RL Unlocks Offline Learning of Generalizable Language-Conditioned Policies with Low-fidelity Data. Proceedings of the 42nd International Conference on Machine Learning, in Proceedings of Machine Learning Research 267:49628-49658 Available from https://proceedings.mlr.press/v267/pouplin25a.html.

Related Material